read.md.qualtrics: Read MaxDiff data for a study object from a Qualtrics QSV...

In cnchapman/choicetools: Tools for Choice Modeling, Conjoint Analysis, and MaxDiff analysis of Best-Worst Surveys

Read MaxDiff data for a study object from a Qualtrics QSV file

Description

This function will bring in MaxDiff data from a Qualtrics QSV file and prepare it for plotting and estimation by other routines in this package. However, Qualtrics's MaxDiff format is brittle and the data must be exported following a very specific regimen. See Chapter 10 of Chapman and Rodden, "Quantitative User Experience Research" for more details, and see the data set in this package, "qualtrics-pizza-maxdiff.csv" for an example.

Usage

read.md.qualtrics(md.define, use.wd = FALSE)

Arguments

md.define	A study object created by parse.md.qualtrics or by hand
use.wd	Whether the filename in parse.md.qualtrics needs to be prepended with a working directory (also in parse.md.qualtrics) as opposed to being a fully specified path (FQDN).

Details

Qualtrics can export data in many different ways, and in general it does not label MaxDiff data in a systematic way. To process such data, we need a standard format to be able to infer the data setup. This should be followed precisely. The requirements are:

1. When fielding the survey, all MaxDiff items must be REQUIRED so the data are complete (respondents should not be able to get away from making difficult tradeoffs; that's part of the choice model process).

2. In Qualtrics, choose to export the data as a CSV.

3. Convert categorical responses (e.g., "best") to numeric (e.g., 2).

4. Tell it include the randomized design order. This is required to estimate the model and parse the data input.

5. Check that those all succeeded and that the data generally have a sparse matrix (most columns are empty), that the "DISPLAY ORDER" columns with the experimental design are present, and that the "best" and "worst" choices are numerically coded.

Be sure to read all output carefully to make sure it matches your expectations. If the import fails, check the above requirements carefully and remove any respondents with incomplete data. Check that the exported data set matches the requirements. Compare the "qualtrics-pizza-maxdiff.csv" data set in this package to see whether your data appear similar in structure.

Value

A list with md.block that contains appropriately formatted data to be inserted back into the md.design study object.

See Also

[parse.md.qualtrics] for details on setting up an md.define study object.

cnchapman/choicetools documentation built on May 28, 2023, 9:14 a.m.